Data from our original study suggest that the quantity of human milk consumed had a marked effect on the incidence of late-onset sepsis and/or necrotizing enterocolitis (NEC), the duration of hospitalization, and the growth of the premature infant. The proposed study will investigate the incidence of late-onset sepsis and/or NEC and duration of hospitalization in both a randomized trial of feeding premature infants fortified human milk supplemented with either fortified pasteurized donor human milk or preterm formula, and in a non-randomized comparison between fortified human milk and preterm formula. One potential mechanism to relate human milk feeding to a lower risk of infection, the maternal titer of pathogen-specific functional antibody, will be measured. The infants will be followed for 3 years to determine the long-term sequelae related to feeding human milk during hospitalization. Because of the failure of many mothers to provide sufficient milk to meet their infants' needs, mothers will be studied to determine the effect of stress reduction methods on milk production. The common nursery practice of skin-to-skin contact between mother and infant will be evaluated for its effect on maternal functional antibody titers and milk production. Infants from the nurseries at Texas Children's Hospital will be enrolled if they are less than 30 weeks gestation, the group with the highest risk of developing late-onset sepsis and/or NEC. We use the term fortified mother's milk to indicate mother's own milk to which a commercial preparation of protein, minerals, vitamins, and glucose polymers is added.